WO2002008722A2

WO2002008722A2 - Direct assessment of relative concentrations of variants of an epitope on a dimeric molecule

Info

Publication number: WO2002008722A2
Application number: PCT/EP2001/007985
Authority: WO
Inventors: Werner Naser
Original assignee: Werner Naser
Priority date: 2000-07-12
Filing date: 2001-07-11
Publication date: 2002-01-31
Also published as: AU2001276395A1; WO2002008722A3

Abstract

The invention relates to methods and devices for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants. The result of measurements according to the present invention is directly or indirectly correlated to the ratio of variants of an epitope carried by the analyte under investigation. The methods and devices described are applied to rather variable samples and yield results essentially independent of variations in the absolute values of the analyte investigated.

Description

Direct Assessment of Relative Concentrations of Variants of an Epitope on a Dimeric Molecule

Field of the invention:

The invention relates to methods and devices for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants. The result of measurements according to the present invention is directly or indirectly correlated to the ratio of variants of an epitope carried by the analyte under investigation. The methods and devices described are preferably applied to rather variable samples and yield results essentially independent of variations in the sample, as caused e.g., by volume effects or diurnal variation.

Background of the invention:

The following description of the background of the invention is provided to aid in understanding the invention but is not admitted to be or describe prior art to the invention.

Diseases of bone, among these osteoporosis, are becoming an increasing burden to society. The total cost in the USA in 1992 of osteoporosis related injuries alone is estimated to be at least USD 10 billion (Riggs L, New England Journal of Medicine, 327:620-627 (1992 )).

Osteoporosis as well as a number of other diseases of bone is characterised by an increased rate of bone loss when compared to the rate of loss in a healthy population. The rate of loss has been shown to be highly correlated to the future fracture risk (Christiansen et al., Prediction of future fracture risk. In: Christiansen et al., eds., Proceedings 1993, Fourth International Symposium on Osteoporosis, Hong Kong, Osteopress Aps 1993; pp. 52-54). Biochemical markers of bone metabolism are an important tool to aid in the diagnosis of diseases which result in and are characterised by changes of bone metabolism, no matter whether the disease results in increased or decreased bone formation or in creased or decreased bone resorption (Delmas, P. D. et al., Journal of Bone and Mineral Research (1986), 1: 333-337). Biomarkers have recently been demonstrated useful to assess efficacy of anti-resorptive treatment (Bjamason, N. H., Christiansen O, Bone 26(6), p. 553-560, 2000)

Due to the fact that osteoporosis is clearly preventable but only partially treatable, the early detection of osteoporosis is crucial if bone mineral content is to be preserved in menopausal adults and bone deterioration is to be prevented early on. Studies have shown that the rate of bone loss after menopause is frequently increased as compared to the rate of loss in pre- menopausal women. Many women in the first years after menopause are losing bone at a rate of greater than 3% and up to 7% per year. Further, in the majority of patients presenting with osteoporosis, 20%-40% of Bone Mineral Content has already been lost before diagnosis is made. Even nowadays osteoporosis is severely under-diagnosed.

A variety of technically sophisticated methods have been developed to assist in predicting the likelihood of bone fractures, such as bone densitometry and quantitative ultrasound procedures. Both of which represent a highly specialised means for measuring bone mineral content. These densitometric measurements provide a static picture and do not give any clue regarding the metabolic events going on in bone tissue. Only the repeated measurements after many months can be used to assess the rate of bone loss. Hence, when these tests are performed, in the majority of cases they only confirm whether a patient has lost significant quantities of bone mineral content or not.

Such densitometric approaches are of limited use in actually diagnosing those peri-menopausal adults who are likely to become osteoporotic, or are at risk to develop osteoporosis. In addition, these procedures also are quite expensive and not always re-imbursed by public health care providers, e.g. in Germany.

For a more successful detection of the onset of osteoporosis the use of serum or urinary concentrations of key biochemical markers has been suggested.

Several recent attempts to diagnose osteoporosis and to monitor bone turnover have focused on the measurement of special collagen degradation products (CDPs). In the past amino acid derivatives, typical for collagen, like hydroxyproline or hydroxylysine have been used. Other methods relying on special cross-linking structures of type I collagen have been described, e.g. in US 5,700,693.

Nowadays much focus clearly is on short peptides, which are released upon or after the degradative action of osteoclasts. These collagen degradation products (=CDPs) comprise the N-terminal and the C-terminal collagen telo-peptides also called NTX or CTX, respectively.

In EP-B-0 394 296, a method for measuring bone resorption is described, based on immunological reagents with specificity to both, the cross-linking structure itself (a pyridinium cross-link) and to the peptide sequence attached to this cross-link. A product based on this approach is FDA-registered and called Osteomark™.

Others (Roche Diagnostics in EP-B-07 1 415, Osteometer in PCT WO 91/13909) have used synthetic peptides as immunogens and found that antibodies to collagen peptides, which do not depend on the presence or absence of a (pyridinium) cross-link) provide also very valuable tools in order to measure CDPs. Whereas Osteomark™ measures amino- or N-terminal CDPs the alternative product Cross Laps™ from Osteometer measures carboxyterminal or C-terminal CDPs.

When measured from urine samples all the above analytical procedures require that correction measures for urine concentration, volume, or dilution are taken. In some instances urine is collected for 24 hours. This however is complicated, not well-accepted by the patient and quite time-consuming. Most commonly creatinine is measured as an indicator for urine concentration and kidney function and the values measured for the above analytes are expressed as relative concentration over creatinine values e.g. per mmol or mg of creatinine.

All the above analytes largely depend on bone mass. The more bone mass - the more organic bone matrix - available the higher the values for any of these analytes. This, however, complicates the diagnosis of osteoporosis with patients having low peak bone mass or having already lost a significant percentage of bone mass.

It was not until quite recently that it has been discovered that some of the C-terminal CDPs contain a fairly unusual β-aspartic acid amino acid linkage (Bonde et al., PCT/WO 96/12193). Such β-linkage is typical for collagen released from "old" bone. Newly formed (and degraded) collagen on the other hand contains the normal or α- linkage of aspartic acid to glycine in the so-called 8AA-peptide of CrossLaps. The ratio of so-called α-CTX (regular peptide bond) to ^β- CTX (β-peptide bond) is useful to further improve the clinical assessment of bone related disorders.

The CDPs summarised as NTX or CTX represent rather small analytes of which most are in the molecular weight range of 1500 to 5000 Dalton. NTX as well as CTX molecules comprise at least one small peptide sequence or peptide epitope, which is subject to extra-cellular chemical rearrangement. Such rearrangement predominantly leads to the above described β-amino acid linkage.

As described in PCT WO 96/12193 the ratio of α-8AA C-terminal CDPs (α-CTX) to the β-8AA C-terminal CDPs (β-CTX) appears to represent a significant improvement for the assessment of bone turnover based on measurement of the ratio of these CDPs. However, both α-CTX as well as β-CTX have to be measured independently. Since both can be measured from the same sample, creatinine correction is not a must. However, again and still, at least two independent measurements are required, both measurements being subject to variations during measurement and subject to assay variations from lot-to-lot of the at least two products used.

High variability of analyte concentrations has been shown for all collagen degradation products. This high variation of marker molecules within a single person or patient represents the most severe drawback in diagnostic use of CDPs markers (cf. e.g., Hannon et al., (1998) Journal of Bone and Mineral Research, 13:1125-1133). It is known that diurnal variation of urinary markers may be as high as ± 50 %. Such high diurnal variation requires extremelx high changes in marker levels in order to obtain clinically significant changes or in order to be of diagnostic significance. This high diurnal variation is extremely critical for all CDP-based products known.

It was the task of the present invention to significantly improve on existing diagnostic procedures in the field of bone metabolism and to overcome some of the critical disadvantages as described above.

Quite surprisingly it could be demonstrated that it is possible to assess the ratio of epitope variants on a dimeric molecule like NTX or CTX directly in a single measurement using the methods and devices according to the present invention. The inventive approach described herein further has the advantage that variable samples like urine may be used without correcting for dilution or volume effects.

The new methods and devices for the assessment of the relative concentrations of epitope variants developed greatly improve on the assessment of bone metabolism and may find broad usage in the field of osteoporosis.

Most strikingly, diurnal variation is not significant for the ratio of analyte variants as determined with the present invention. Thus diagnostic utility of CDP-based data is largely improved.

The above described problems encountered with state-of-the-art procedures have been solved based on a new technical approach as described in the description section by methods and devices as claimed.

Summary of the invention

The invention relates to methods and devices for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants. The result of measurements according to the present invention is directly or indirectly correlated to the ratio of variants of an epitope carried by the analyte under investigation. The invention also relates to the assessment of collagen telo-peptide variants by aid of only one variant-specific binding partner.

Some collagen degradation products, especially the collagen telo-peptides called NTX and CTX have been found to comprise several variants of otherwise unique epitopes. These molecules comprise a cross-link structure derived from two or three lysine or hydroxylysine residues as well as short amino acid stretches of type I collagen. For both, NTX and CTX, it is known that the collagen type I sequences contained can undergo a rearrangement of the aspartic acid to glycine peptide bond to form a so-called β-linkage. Since the β-linkage is typical for "old" bone the ratio of fragments carrying either the regular α-linkage or the β-linkage can be used to assess bone metabolism or bone balance.

Classical methods yield absolute values for biochemical bone parameters. No single value in itself can be used to assess whether more bone is newly formed or whether more old bone is degraded, thus the bone balance, i.e. information on whether the net result of the processes is bone formation or bone resorption or loss can not be obtained. Bone balance can however, be assessed by analysing ratios of analytes, which are indicative for bone formation on the one hand and for bone resorption on the other hand. All such methods require independent measurement of at least two different analytes or independent measurement of at least two variants of one epitope on such an analyte.

It has now surprisingly been found that it is possible to assess bone metabolism in a single measurement. Highly variable sample sources, e.g. urine may be used for such assessment. Using the methods and devices of the invention it is not necessary to undertake correction measures for sample variables like concentration, dilution or volume, nor to use timed sampling. At most, measures have to be taken to ensure that enough analyte is present.

The invention comprises methods for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants, characterised in that,

(a) a first variant-specific binding partner

(b) a sample containing said first variant of said epitope in excess as compared to said first binding partner are incubated under conditions allowing for binding between said epitope variant and said first binding partner

(c) complexes between said first binding partner and said analyte are formed wherein the ratio of variants in said complexes correlates to the ratio of said variants in said sample

(d) a detectably labelled second variant-specific binding partner is used to detect the fraction of complexes (c) carrying the epitope variant recognised by said second binding partner

(e) the signal generated (d) is correlated to a calibration or standard curve

Optionally a step may be included to ensure that enough analyte is present in the sample volume analysed.

With the methods and devices of the present invention only single measurements are required. The results are directly or indirectly correlated to the relative concentration of epitope variants. In the case of type I collagen telo-peptides the relative concentrations of certain variants of telo- peptide epitopes represent most appropriate means to assess bone metabolism. Such methods and devices are used to assess net bone balance, efficacy of treatment regimens which are aimed at changing bone metabolism, and are very useful in diagnosis of osteoporosis.

The methods and devices of the present invention feature an additional striking advantage: they can be used with highly variable samples and no measure have to be taken to correct for volume or concentration effects. This highly desirable advantage is brought about by the fact that not absolute concentrations of analytes or of specific variants of these analytes are measured or required, but rather relative concentrations of variants of analytes are measured.

Such relative concentrations of epitope variants or the ratio of one variant of an epitope on the analyte to another variant of this epitope on the analyte are largely independent of any volume effects.

Diurnal variation of biomarkers measured from urine is one of the major obstacles to broad use of such markers. Surprisingly it has now been found and demonstrated that diurnal variation is not significant or may be not even existent for the relative concentrations of variants of epitopes on dimeric collagen degradation products. This surprising finding opens new possibilities for broad use of CDP-variant ratios in many questions and fields of collagen disorders, e.g., in facilitating risk assessment or screening for osteoporosis.

Detailed description of the invention

Diseases correlated to changes or abnormalities in collagen structure or metabolism include Paget's disease, osteogenisis imperfecta, tumour metastasis, dwarfism, rheumatoid arthritis , osteo-arthritis, as well as osteoporosis.

The field of osteoporosis and its diagnosis e.g. from urinary samples may be considered to be one of the most advanced - though still a quite controversy area - of "quantitative" diagnosis using urine as a sample. In recent years a lot of different attempts have been undertaken to diagnose osteoporosis through analysis of urinary samples. The field of osteoporosis as well as the tremendous advantages of the present invention shall be described in some detail.

Diseases of bone, among these osteoporosis, are becoming an increasing burden to society. The total cost in the USA in 1992 of osteoporosis related injuries alone is estimated to be at least USD 10 billion (Riggs, New England Journal of Medicine, 327:620-627 (1992 )). Definition as well as diagnosis of osteoporosis is complicated by the fact that the clinical manifestation of the disease due so-called osteoporotic fractures is seen many many years after the actual onset of the physiological processes leading to it. Nowadays definitions for osteoporosis tend to include the physiological basis of the disease and to define osteoporosis as a disease characterised by a negative bone balance or as a disease characterised by an increased rate of bone resorption or an increased rate of (bone) loss.

Due to the fact that osteoporosis is clearly preventable but only partially treatable, the early detection of osteoporosis is crucial if bone mineral content is to be preserved in menopausal adults and bone deterioration is to be prevented early on.

Studies have shown that the rate of bone loss after menopause is frequently increased as compared to the rate of loss in pre-menopausal women. Many women in the first years after menopause are losing bone at a rate of greater than 3% and up to 7% per year. Further, in the majority of patients presenting with osteoporosis, 20%-40% of Bone Mineral Content has already been lost before diagnosis is made.

Such densitometric approaches are of limited use in actually diagnosing those perirnenopausal adults who are likely to become osteoporotic, or are at risk to develop osteoporosis. From a single point measurement it is not possible to predict, whether bone loss will be rapid or insignificant in the future. In addition, these procedures also are quite expensive and not always reimbursed by public health care providers, e.g. in Germany.

For a more successful detection of the onset of osteoporosis the use of serum or urinary concentrations of one or several key biochemical markers has been suggested (cf.: Delmas, P. D. et al., Journal of Bone and Mineral Research (1986), 1: 333-337). This suggestion is based on findings that osteoporotic processes may translate to changes in biochemical markers which can be measured e.g. from serum or urine.

A large variety of such markers, like bone alkaline phosphatase, osteocalcin, tartrate resistant acidic phosphatase, etc., is known to the expert in the field, and must not be discussed here in detail. It shall however be iterated, that these markers have not found broad use e.g. in assessing osteoporosis.

Most of the more recent approaches to diagnose osteoporosis from samples like serum or urine by aid of biochemical markers are subject to patent applications and several of them shall be mentioned in some detail.

PCT/WO 96/04544 is entitled "Urinary Test Strip for Determining Calcium Loss". It describes at length the problems encountered when using urine as a sample. The amount of calcium excreted into urine is directly related to bone turn-over. And, under the proviso, that no other factors, like diet, contribute significantly to calcium excretion, total urinary calcium is related to overall bone turn-over and bone loss. However, either 24h urine samples have to be collected and the amount of calcium excreted per 24 hours has to be determined or, its excretion has to be matched against the known constant excretion product of urinary creatinine. The ratio between calcium and creatinine is said to be applicable to the detection of osteoporosis.

PCT/WO 96/04544 teaches the use of test strips comprising means for two independent measurements, one for measurement of calcium and one for measurement of creatinine. Both molecules are measured independently based on the calcium- and on the creatinine- specific colour developed in the respective area of the test strip. Colours generated can be read visually or by photometric devices and are used as basis for the calculation of the calcium to creatinine ratio. The improvement of such a device being that both independent reactions are performed on the same test strip, as compared to the conventional approach of measuring both molecules on clinical-chemical analysers and thereafter calculating the ratios.

Whereas PCT/WO 96/04554 measures an inorganic component of bone (calcium) the most advanced technologies to assess bone metabolism, e.g., to detect osteoporotic conditions the efficacy of treatment regimens or the presence of bone metastasis are based on the measurement of organic material released from bone during bone turn-over. Most attempts to diagnose osteoporosis and to monitor bone turnover have focused on the measurement of special structures derived from collagen degradation.

The first marker molecules investigated for their diagnostic potential have been amino acid derivatives, typical for collagen, like hydroxyproline or hydroxylysine or its glycosides. Detection and use of hydroxyproline is described in US 3,600,132. Hydroxylysine and its derivatives are discussed in Krane, S. M. and Simon, L. S., Develop. Biochem. 22: 185 (1981).

Other methods relying on special cross-linking structures typical of type I collagen have been described, e.g. in US 5,700,693. These methods are based on detection of the coss-link structures (e.g. the pyridium or the deoxy-pyridinium cross-link) themselves.

Nowadays much focus clearly is on short peptides, which are released upon or after the degradative action of osteoclasts.

In EP-B-0 394296, a method for measuring bone resorption is described, based on immunological reagents with specificity to both, the cross-linking structure itself (a pyridinium cross-link) and to the peptide sequence attached to this cross-link. A product based on this approach is FDA-registered and called Osteomark™.

Others (Roche Diagnostics in EP-B-0711 415, Osteometer in PCT/WO 91/13909) have used synthetic peptides as immunogens and found that antibodies to collagen peptides, which do not depend on the presence or absence of a (pyridinium) cross-link) provide also very valuable tools in order to measure CDPs. Whereas Osteomark™ measures amino- or N-terminal CDPs the alternative product CrossLaps™ from Osteometer measures carboxy-terminal or C-terminal CPDs.

It was not until quite recently that it has been discovered that some of the C-terminal CDPs contain a fairly unusual β-aspartic amino acid linkage (Bonde et al., PCT/WO 96/12193). Such ^β-linkage is typical for collagen released from "old" bone. Newly formed (and degraded) collagen on the other hand contains the normal linkage of aspartic acid to glycine in the so- called 8AA-peptide of CrossLaps. The details of the above mentioned patents disclosing techniques for measurement of CDPs are herewith included by reference. The skilled artisan will find in the related patent families the technical advice required for production of specific antibodies as well as for obtaining naturally occurring or synthetically produced analytes or analyte analogues. Especially the synthetic peptides representing CDP-epitopes as disclosed in EP-B-711 415, WO 95/08115, PCT/WO 96/36645 are useful to generate antibodies which may be used in a device or procedures according to the present invention.

Information of a more general nature on how to synthesise, screen for and to use synthetic peptides as well as so-called mimetics thereof can be found in WO 91/13909 + WO 95/20 764. The methods disclosed there are included by reference. Such mimetics or analogues of analytes, epitopes, or variants of epitopes may be used to modify (as required) the binding/release properties of such variant molecules to/from their specific binding partner.

The vast majority of the above discussed markers aiming at analysis of bone metabolism is measured from urine. With urine as sample, all these methods require that a) the analyte is determined, that b) independently and by quite a different procedure, creatinine is determined and that c) the results are expressed as CDP/creatinine ratios, unless 24 hour urine collections would be available. It is only the CDP/creatinine ratios, which allow for meaningful comparisons of data from one measurement to the other in and between individuals or patient groups.

Independent from what has already been said for kidney function itself, it is obvious, that the excretion of CDPs (or their presence in blood) must be related to both a) the actual bone mass and b) to the metabolic activity of either osteoblasts, osteoclasts or both. As a consequence thereof high bone turn-over at moderate to high bone mass will result in high to very high CDP/creatinine ratios.

Low bone turn-over on the other hand will yield low to very low CDP/creatinine ratios at moderate as well as at low overall bone mass. Diagnosis, however, is far less clear with other patient groups. Most critically, a patient already presenting with a low bone mass and a moderate to high bone-turnover osteoporosis may not be adequately detected, respectively diagnosed, by CDP/creatinine-measurements. Despite the proven general usefulness of CDP/creatinine-ratios such false negative results represent a significant problem and further improvements are required.

As described the ratio of so-called α-8AA C-terminal CDPs (α-CTX) to the so-called β-8AA C- terminal CDPs (β-CTX) appear to represent a significant improvement for the assessment of bone turnover based on measurement of CDPs. (The terms α-8AA and β-8AA, α-CTX and β- CTX will be described in detail below). This ratio has been suggested to correlate to bone balance or the rate of bone loss. According to PCT WO 96/12193, both α-CTX as well as β-CTX have to be measured independently. Since both can be measured from the same sample, creatinine correction is not a must. However, again and still, at least two independent measurements are required, both measurements being subject to variations during measurement and subject to assay variations from lot-to-lot of the at least two products used. This involves technical problems, handling problems, handling time and costs for two independent and not directly interrelated measurements.

Somewhat in analogy to PCT WO 96/04554 (the calcium and creatinine test strip, described above) a device for the simultaneous measurement of both, the concentration of Osteomark™ as well as the concentration of creatinine has quite recently been developed. J. Blatt et al., describe such a device in Clinical Chemistry 44;9,p2051+2052, 1998. This article is entitled "A Miniaturised, Self-contained, Single-use, Disposable Assay Device for the Quantitative Determination of the Bone Resorption Marker NTX in Urine". This (single-use!) handhold device also contains all the equipment necessary to measure both the signals developed by the NTX- specific reaction as well as by the creatinine determination, to calculate the ratios of NTX/creatinine and to display the rationed result.

It has now surprisingly been found that it is possible to overcome most of the problems described above as encountered with state of the art procedures. The methods and devices according to the present invention largely improve on any diagnosis of urinary analytes by directly correcting for sample dilution in one and the same measurement. An independent measurement of creatinine as a reference molecule or another analyte or variant thereof is no longer required.

Most surprisingly it has been found that it is possible to directly assess the rate of bone turnover, which is correlated to net bone balance or to the rate of bone gain or bone loss, in one single measurement.

Whereas the high clinical relevance of biochemical markers for groups of patients, e.g. in clinical studies is broadly accepted, there is the severe drawback that the values measured for one individual patient are of much less clinical significance. Only quite recently it has recently been demonstrated by Bjarnason, N. H., Christiansen O, Bone 26(6), p. 553-560, 2000, that biochemical markers might be useful to monitor in an individual patient the response (or failure to respond) to hormone replacement therapy. The rate of (bone) loss has been shown to be highly correlated to the future fracture risk (Christiansen et al., Prediction of future fracture risk. In: Christiansen et al., eds., Proceedings 1993, Fourth International Symposium on Osteoporosis, Hong Kong, Osteopress Aps 1993; pp. 52-54, Delmas, P. D., Bone 13: 517-521 , 1992). Therefore the rate of loss is an important tool to aid the diagnosis of diseases which result in changes of bone metabolism, no matter whether this results in increased or decreased bone formation or increased or decreased bone resorption.

Surprisingly it has now been found that it is possible to assess the rate of (bone) loss by measurement of relative concentrations of variants of collagen telo-peptides. In more general terms the relative concentrations of these variants are determined by ajnethod for the assessment of the ratio of variants of an epitope on an analyte molecule in a sample, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants, characterised in that,

(a) a first variant-specific binding partner

(b) a sample containing said first variant of said epitope in excess as compared to said first binding partner are incubated under conditions allowing for binding between said epitope variant and said first binding partner (c) complexes between said first binding partner and said analyte are formed wherein the ratio of variants in said complexes correlates to the ratio of said variants in said sample

(e) the signal generated (d) is correlated to a calibration or standard curve. It is good and general practise that such calibration or standard curves preferably are developed and established in independent experiments and results measured are extrapolated from such data or curves.

In order to understand and appreciate the progress made it is necessary to describe some features of bone collagen and bone metabolism in more detail.

Type I collagen is predominantly found in bone and present there in large amounts. It is especially this molecule which has gained a lot of interest in the field of bone metabolism, e.g. in the diagnosis of bone metastasis, monitoring the efficacy of treatment or in diagnosis of osteoporosis is type I collagen. About 90% of the organic material found in the extra-cellular matrix of bone is type I collagen. During (physiological or pathological) bone turn-over osteoclasts resorp bone matrix, thereby releasing collagen degradation products (CDPs) and forming so-called resorption lacunes. Osteoblast or osteoblast (precursors) attach to the resorption sites and new bone matrix is formed over time. In healthy adults the resorption and the formation of bone are in equilibrium and bone mass stays constant.

Osteoporosis, especially postmenopausal osteoporosis, especially in the first ten to fifteen years after menopause is characterised by a negative bone-balance, i.e. more bone is resorped that formed. This results in the above mentioned loss of bone and eventually in the disease called osteoporosis. The rate of loss measured and expressed in terms of bone mineral density may be as high as five to seven percent in a single year.

Osteoporosis as well as a number of other diseases of bone is characterised by an increased rate of bone loss when compared to the rate of loss in a healthy population. Other clinical manifestations, e.g. Morbus Paget are characterised by net bone gain. Obviously growth during childhood is also characterised by net bone gain or a positive bone balance.

The type I collagen molecule, like type II and type III collagens is formed in the organism as a procollagen molecule, comprising N- and C-terminal propeptide sequences attached to the core of the molecule. After removal of the propeptide sequences three collagen core monomers form a triple-helical collagen fibril. The core molecule itself consists of a central helical part taking part in formation of the triple helix and contains at both ends linear or non-helical stretches. These linear, non-helical sequences are also called telo-peptide sequences. The collagen molecules, produced and secreted by the osteoblast are subject to intra- and extra-cellular posttranslational modifications. Especially, the N- and/or C-terminal portions of collagen are subject to extracellular, intra- and intermolecular cross-linking. They have an important function as sites of intermolecular cross-linking of collagen fibrils. The intermolecular cross-links provide biomechanical stability to the collagen fibrils.

Di or tri-valent cross-links originating from two or three lysine or hydroxylysine residues are most prominent. The chemical structure of these cross-links is known to the skilled artisan, details can e.g. be found in some of the patent literature discussed above (US 5,700,693; EP 0 394 296). Detailed knowledge upon the (chemical) structure of these cross-links, however, is not relevant to the present invention. Any structure linking at least two telo-peptide sequences, each comprising an epitope, together, may be part of the analyte assessed. Relevant for the embodiments according to this invention is the fact that as a result of bone resorption small collagen peptide residues, the so-called telo-peptides, from the N-or C-terminal portions of the collagen molecule - still containing a cross-linking structure - are found in body fluids. The CDPs summarised as NTX or CTX represent rather small analytes most of which are in the molecular weight range of 1500 to 5000 Dalton. NTX as well as CTX molecules comprise at least one small peptide sequence or peptide epitope, which is subject to extra-cellular chemical rearrangement. Such rearrangement leads to epitope variants which are of high diagnostic relevance, as biological correlates for bone formation or bone resorption.

In the case of the C-terminal telo-peptides most of the fragments contain twice a sequence comprising eight amino acids, the so-called 8-AA sequence Glu-Lys*-Ala-His-Asp**-Gly-Gly-Arg (Lys* may be part of a cross-linking structure; Asp** may be linked to Gly by regular peptide linkage (=α-8AA) or by an iso-peptide bond (=β-8AA) (Fledelius et al., The Journal of Biol. Chem. 272: 15, 9755-9763, 1997). N-terminal telo-peptides have been found to comprise one α1- and one α2-sequence of collagen type I or two α2-sequences of collagen type I (N-telo- peptide sequence of type I collagen (α1): Asp-Glu-Lys-Ser-Thr-Gly-Gly; N-telo-peptide sequence of type I collagen (α2): Gln-Tyr-Asp*-Gly-Lys-Gly-Val-Gly) (Hanson et al., 1992). Collagen telo-peptides which are slightly smaller or larger or additionally carrying sugar residues are known for both NTX and CTX. For sake of convenience only the most prevalent forms, as described above, are discussed further-on.

Preferred analytes are molecules, especially small molecules, comprising at least two epitopes in an at least dimeric molecule. At least one of these epitopes being present in at least two variants. Also preferred are analytes wherein both epitopes on the at least dimeric molecule may be present in form of at least two variants. The above telo-peptides represent prototype analytes to which the present invention can be applied.

Such analyte molecules in most cases and most preferred will be the result of intermolecular cross-linking. Well-known examples are the cross-linked intermediate filament proteins, like elastin or collagens, especially type I, II and III collagens. Preferred analytes are degradation products comprising at least two epitopes on an at least dimeric structure which are present in form of degradation products. Such degradation products preferably have a molecular weight of less than 20000 D, more preferred of less than 10000 D, and even more preferred of less than 5000 D. Preferred analytes in the sense of the present invention are especially collagen telo-peptides, especially the collagen telo-peptides derived from type I type II or type III collagen degradation. Type II and type I collagen telo-peptides are even more preferred, most preferred are collagen telo-peptides derived from type I collagen. Collagen telo-peptides in one embodiment comprise twice the same epitope. Special examples of type I collagen telo-peptides are CTX-molecules comprising twice the 8AA-sequence and those NTX-molecules, comprising twice the N-telo- peptide α2-sequence. In a further embodiment N-telo-peptides comprising one α1- and one α2- sequence of collagen type I are covered. The relative concentrations of variants of epitopes on these collagen degradation products are used in order to assess bone metabolism, especially to assess bone balance, rate of bone loss, or to aid in diagnosing osteoporosis.

As indicated above, the peptide bond between the amino acids aspartic acid and glycine is subject to a statistical chemical modification leading to a so-called isomerised peptide bond also termed iso- or β-peptide bond or β-linkage. This isomerisation takes several weeks to months. Therefore especially "old" bone contains significant amounts of such a beta-linkage, which when released from bone is indicative for bone resorption. On the contrary, telo-peptide fragments containing the regular peptide in α-linkage are more indicative for turn-over of newly or recently synthesised bone matrix, which in turn, when released from bone and found and body fluids, is considered representative for bone formation.

For completeness it shall be mentioned that the isomerisation described may lead to additional distinct peptide bonds termed αD and βD. Details on this topic are found in PCT/EP 97/04372.

With other words, collagen telo-peptides may contain epitopes within one and the same type of analyte either indicative for processes of bone formation or indicative for processes of bone resorption.

Most CTX fragments are either composed of twice α-8AA, twice β-8AA or/one each α-8AA and β-8AA. This mixture is statistical and is best described by the formula (a+b)² = a² + 2ab + b² (a and b each representing one variant of the epitope under investigation).

The statistical mixture of epitope variants on CTX-molecules is quite critical in quantitative immuno assays according to the sandwich assay principle, i.e., when a first and a second specific binding partner are used. To illustrate the problem, a sandwich immuno assay may be designed to measure CTX fragments comprising twice β-8AA. The sample however, does not only contain this molecule but as well CTX-fragments comprising one variant each of α-8AA and β-8AA, and most critically also molecules with one variant each. This molecule will bind to the first or to the second specific binding partner. However, sandwich formation is not possible due to the second variant being α-8AA. Such molecules compete with the correct analyte (twice β- 8AA) for the binding sides on the specific binding partners and thus interfere in the assay. Such interference may not be critical as long as both specific binding partners are present in excess. This for example is true in most cases when serum is used as sample.

Concentrations of CDPs in urine in many cases are about 50 to 100-fold higher as compared to serum. If not properly diluted, such samples must interfere in a sandwich assay for CTX comprising twice β-8AA due to the presence of CTX-fragments comprising one each of α-8AA and β-8AA. This results in false measurements.

It has now surprisingly been found that this interference by the mixed variant analyte molecules can be used to great advantage in order to assess the relative concentrations or ratios of epitope variants in a single measurement.

This great progress is illustrated by aid of three theoretical examples of the following table (both variant-specific reagents reactive to βL-epitope variants).

As can be seen from the above table, changes in relative concentrations of epitope variants (0.3 to 0.7) translate into quite significant changes as measured (18 % triple to 54 %). Epitope in the sense of the present invention refers to a structure recognised by a specific binding partner, e.g. a lectin, a monospecific antibody, or a monoclonal antibody. Peptide epitopes usually comprise four to eight, especially five to seven amino acids. E.g., the α2- peptide sequence of the N-terminal telo-peptide and the 8AA sequence of the C-terminal telo- petide each represent an epitope. These epitopes can be present in form of several variants, the regular αL-form the quite prominent βL-form, or the rare αD- and βD-form.

Variant of an epitope in sense of the present invention is used to describe a secondary modification of an epitope. Well-known secondary modifications are e.g. the addition of sugar residues via specific glycosylation reactions or due to diabetic complications (resulting in structures known as advanced glycosylation end-products), phosphorylation of amino acids, cross-linking via sulfhydryl groups, lysine side chains, or due to action of transpeptidases and the above described isomerisation of peptide bonds. Such modification must be detectable by aid of a specific binding partner.

A specific binding partner for a variant of an epitope does preferentially react with the one variant to which it is specific and exhibits little or insignificant cross-reactivity to other variants of that epitope. The amount of cross-reactivity which might be tolerable in order to carry out the invention depends on the relative concentration of the variants investigated. Preferred are variant-specific binding partners with less than 10% cross-reactivity, more preferred with less than 3% cross-reactivity and most preferred with less than 1 % cross-reactivity. Variant-specific binding partners preferably are antibodies, however, any other structure bringing about a variant-specific binding, e.g. a lectin, a receptor, or a synthetic binding structure may be used.

Variant-specific antibodies to epitopes on collagen telo-peptides can be produced according to procedures known in the art, e.g. as described for the several variants of CTX-epitopes, the regular αL-form the quite prominent βL-form, or the rare αD- and βD-form in PCT/EP 97/04372. The same holds true for the α2-sequence of N-terminal telo-peptides. References (e.g., EP-B- 0711 415; PCT/WO 99/00668, PCT/WO 95/08115 or PCT/WO 96/12193 describing the production of antibodies to collagen telo-peptides are broadly available and procedures described there herewith included by reference. Specificity for epitope variants is established according to standard procedures, e.g. by competition experiments using the specific variant of an epitope as well as other variants as competitors. Specific detection of CTX-molecules containing twice β-8AA-variant of the 8AA-epitope is possible and has been described in WO 98/26286. Monoclonal antibodies generated according to the procedures given in EP-B-0 711 415; PCT WO 95/081 5 or PCT/WO 96/12193 are available or can be generated which can be used to specifically detect each α-8AA x α-8AA- CTX, α-8AA x β-8AA-CTX and β-8AA x β-8AA-CTX in a sandwich immunoassay.

As described in WO 98/26286 the monoclonal antibodies 1103 and F12 bind only to one β-8AA chain each of the CTX-molecule and thus are capable of forming a sandwich comprising a first variant-specific antibody, the β-8AA/β-8AA-CTX and a second variant-specific antibody. In cases where the antibody used would be capable of binding to both its variants on the analyte, single binding structures, most preferred, so-called F(ab)-fragments are preferentially used.

The methods and devices according to the present invention are used to assess the ratio of variants of an epitope on an analyte molecule in a sample, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants. It is also made possible according to this invention to assess the relative concentration of variants of collagen telo-peptides in a sample, by aid of a first binding partner specific for a type I collagen telo-peptide sequence and one binding partner specific to an epitope variant of these telo- peptides.

Incubation conditions are selected as appropriate. The selection of buffers and buffer additives is well-known to the skilled artisan and can be found in text-books dealing with quantitative and qualitative immuno assays (e.g., Tijssen, 1995, supra).

The method according to the present invention require that a first specific binding partner is made available in limited amounts and reacted with a sample comprising the variant of an epitope recognised by said first binding partner in excess.

Complexes are formed between the first specific binding partner and the analyte molecules carrying the first (epitope) variant at least once. Under the above conditions the ratio of epitope variants bound via the first specific binding partner is correlated to the ratio actually present in the sample. More specifically all binding sides will occupied by the first epitope variant and the molecules bound will either carry the same variant or another and interfering variant. If, e.g., the variants a and b are present in equal amounts they will be present on a di-meric molecule with 25% as a², 50% as 2ab and 25% are b². If the first binding partner is specific for variant b only the 2ab and b² molecules are bound. This means that about 67% of binding sides on the first specific binding partner will be in form of a complex comprising the 2ab analytes and about 33% in turn will comprise b² molecules. Either fraction of these complexes between the first binding partner and the analyte can be detected by aid of a second variant-specific binding partner. The second variant-specific binding partner may also be called detection reagent.

With more variants of the epitope under investigation present in significant amounts, only statistics are more complicated. The method itself, however, is working the same way and ratio of two of such multiple variants can as well be easily determined.

The detection itself in the above example is performed either by a binding partner specific for variant a or by a specific binding partner for variant b.

The complex-formation between analyte molecules and first specific binding partners is over a wide range not dependent on the concentration or dilution of the sample as long as the first specific binding partner present is the limiting factor.

The fraction of complexes carrying the epitope variant of interest can be detected by aid of a binding partner which is specific for this epitope variant. Such detection and evaluation of the results can be performed according to standard procedures.

It is preferred that the second specific binding partner is detectably labelled. Such labelling may be direct or as well indirect. Procedures and labels for direct and indirect labelling are well- known to the skilled artisan and are e.g. described in Tijssen, 1995, supra, or J.Beesley, "Colloidal gold: A New Perspective for Cytochemical Marking", Microscopy handbooks 17, Oxford University Press, 1989.

Some appropriate labels e.g. are enzymes, coloured latex particles, fluorescent or chemiluminescent labels, or colloidal metals, like colloidal gold shall be specifically mentioned. Coloured latex particles and colloidal gold particles are preferred labels. Of course it is also possible to detect the second variant-specific binding partner by using a binding partner for this molecule which does not bind to the first specific binding partner.

Various test designs are possible to carry out the methods according to the present invention.

With concentrated samples, i.e., analyte present in excess as compared to the first specific binding partner, the assay may be performed according to standard sandwich immuno assay principles e.g. using microtitre plates, coated beads or tubes to bind the first variant-specific binding partner. Such assays may e.g. make use of first or second morning void urine samples, which are known to contain high amounts of CDPs.

In a preferred embodiment the ratio of CDP-variants is determined in a sandwich immunoassay e.g., in microtitre plates or an appropriate clinical analysers. The first variant-specific binding partner is either directly coated to the solid phase in limited amounts or indirectly coated making use of a specific binding pair, most preferred streptavidin-biotin. According to state-of-the-art sandwich procedures it is always the goal to reach a plateau, i.e. to offer this binding partner in such high (saturation) amounts (often above 1μg/ml) that all or at least the majority of analyte molecules is bound, the strategy according to the invention is very different. The first specific binding partner is offered only in limited amount. In case of a streptavidin-biotin system, preferably soultions of less than 100ng/ml of biotinylated first binding partner, more preferred less than 50ng/ml and most preferred 25 ng/ml or less are used to coat the solid phase. During analyses, only a fraction of the analyte present is bound. This is quite different from standard immunological procedures which are designed to measure all analyte present.

In many standard wet chemistry assay set ups it is possible to wash away any unbound analyte after the first incubation step, i.e. after formation of a complex between first binding partner and analyte. In this case it is easy to control that enough second binding partner is added, allowing for labelling and detection of the variant of interest present in the complexes between first binding partner and analyte.

If no washing step is used, care has to be taken to add the second and variant-specific binding partner in excess. The amount required is empirically determined.

The sequence of addition of the individual components taking part in the sandwich formation may be varied as appropriate. If washing is not possible care has also to be taken that binding of the first specific binding partner and of the second specific binding partner do not interfere with each other. Interference could occur if both binding partners are reactive with the same variant of an epitope. In this case reaction of excess analyte molecules with the first variant- specific binding partner - which is preferentially monomeric or per se capable of binding to only one of its variants on a dimeric molecule - has to take place before the excess of detection reagent can be added. It has also been found that it is possible to assess the relative concentration of variants of collagen telo-peptides in a sample, by aid of a first binding partner specific for a type I collagen telo-peptide sequence and one binding partner specific to an epitope variant of these telo- peptides comprising incubating under favourable conditions (a) said first binding partner specific for a type I collagen telo-peptide sequence

(b) a sample containing said collagen telo-peptide sequence in excess as compared to said first binding partner

(c) complexes between said first binding partner and said analyte are formed wherein the ratio of variants in said complexes correlates to the ratio of said variants in said sample (d) a detectably labelled variant-specific second binding partner is used to detect the fraction of complexes (c) carrying the epitope variant recognised by said second binding partner (e) the signal generated (d) is correlated to a calibration or standard curve.

As indicated above it is possible to assess the relative concentration of variants of collagen telo- peptides in a sample, by aid of a first binding partner specific for a type I collagen telo-peptide sequence and one binding partner specific to an epitope variant of these telo-peptides. In case of C-terminal telo-peptides an antibody reactive to both major epitope variants the αL-8AA- sequence and the βL-8AA-sequence, may be used as a first specific binding partner and the fraction of either the αL-8AA-sequence or the βL-8AA-sequence bound is determined by a variant-specific second binding partner as described. In case of N-terminal telo-peptides it is possible to use an antibody specific to the telo-peptide sequence derived from the N-telo- peptide sequence of the α1 -chain of type I collagen as first specific binding partner. The αL- or the βL-sequence variants of the α2-chain of type I collagen are then assessed by binding partners specific for one of these variants.

It is also possible to carry out the present invention with samples suspected of containing the analyte under investigation. Such samples may contain highly variable and even rather low concentrations of the analyte. With such samples it is necessary that measures are taken to ensure that said first variant of said epitope is available in excess as compared to said first variant-specific binding partner.

If the sample is not known to automatically contain high concentrations of the analyte, especially of the epitope recognised by the first specific binding partner, measures are taken to ensure that enough samples is applied and that said epitope is present in excess. ln a further embodiment the invention is represented by a method for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants, characterised in that, (a) a first variant-specific binding partner

(b) a sample suspected of containing said first variant of said epitope are incubated under conditions allowing for binding between said epitope variant and said first binding partner

(c) measures are taken to ensure that said first variant of said epitope is available in excess as compared to said first variant-specific binding partner (d) complexes between said first binding partner and said analyte are formed wherein the ratio of variants in said complexes correlates to the ratio of said variants in said sample

(e) a detectably labelled second variant-specific binding partner is used to detect the fraction of complexes (d) carrying the epitope variant recognised by said second binding partner

(f) the signal generated (e) is correlated to a calibration or standard curve.

The measures to be taken to ensure that said first variant of said epitope is available in excess as compared to said first variant-specific binding partner may be quite different.

In case standard wet chemistry immuno assay formats are used, to which a fixed and limited amount of sample is applied, it has to be ensured by independent procedures, e.g. by assessing density of urine, creatinine content or other reference molecules that enough analyte (variants binding to the first specific binding partner) are present in the sample to be analysed. It is also possible to directly measure the first epitope-variant independently.

In a preferred embodiment the first and the second variant-specific binding partner recognise different variants of said epitope. As described above, most collagen telo-petides comprise twice the same epitope (α2 sequence for NTX or 8AA for CTX) which are mainly present as regular αL- or as isomerised βL-peptide. In case the two variant-specific binding partners recognise different epitopes and are capable of forming a sandwich complex either one can be used as capture or detection reagent.

In another embodiment the methods according to the present invention are carried out using a first and a second variant-specific binding partner which both recognise the same variant of said epitope. As described in WO 98/26286, the specific detection of CTX-molecules containing twice the β-8AA-variant of the 8AA-epitope is possible. The monoclonal antibodies Λ 103 and F12 bind only to one β-8AA chain each of the CTX-molecule and thus are capable of forming a sandwich comprising a first variant-specific antibody, the β-8AA x β-8AA-CTX and a second variant-specific antibody. These reagents can be directly used in methods and devices according to the present invention. In cases where the antibody used is capable of binding to both its variants on one and the same analyte no sandwich formation is possible. Such antibodies can be modifies to yield single binding structures. Most preferred, so-called F(ab)- fragments are used.

In a further preferred embodiment the first and the second variant-specific binding partner used in a method according to the present invention are antibodies. It is well-established that antibodies belong to the most specific binding partners known in nature. Since it is possible to produce variant-specific antibodies. It is preferred that both variant-specific reagents are antibodies, and even more preferred monoclonal antibodies.

Methods for preparation of monoclonal antibodies are well-known in the art, e.g. Campell, A. M., Laboratory Techniques in Biochemistry and Molecular Biology, Vol. 12 (1986). Monoclonal antibodies with specificity for epitope variants on collagen telo-peptides are available or can be generated e.g. as described in EP-B-0 711 415; PCT/WO 95/08115 or PCT/WO 96/12193. As described in WO 98/26286, the specific detection of CTX-molecules containing twice the β-8AA- variant of the 8AA-epitope is possible. The monoclonal antibodies 1103 and F12 bind only to one β-8AA chain each of the CTX-molecule and thus are capable of forming a sandwich comprising a first variant-specific antibody (1103), the β-8AA/β-8AA-CTX and a second variant- specific antibody (F12). It is therefore a preferred embodiment to use monoclonal antibodies in a method according to the present invention as the first and as the second variant-specific binding partner.

In a further preferred embodiment the analyte investigated by a method according to the present invention is a collagen degradation product. Collagen degradation products, especially the so-called telo-peptides represent rather unique molecules. They are derived from the degradation of a formerly coss-linked big molecule. They are degraded to an extend allowing for their secretion into urine. Collagen-telo-peptides belong to the smallest molecules for which sandwich immuno assays have been described. The relative concentrations of variants of these molecules can easily be analysed using a method or a device according to the present invention. Results of such analysis which reflect bone metabolism, especially bone balance are of great clinical and practical utility.

It has been shown (see above) that the collagen degradation product summarised as N- terminal telo-peptides or C-terminal telo-peptides both comprise peptide epitopes which are present in form of several epitope variants. It is a preferred embodiment of the present invention to analyse the ratio of epitope variants for N-telo-peptide(s) or for C-telopeptide(s) in a method according to the present invention.

Since the "regular" peptide comprising the usual α-peptide linkage is known to represent processes related to bone synthesis or bone formation, whereas all other variants are more representative for old bone and processes of bone resorption, both these structures are ideal to investigate the processes of bone formation and bone resorption by aid of one type of molecule.

According to methods of the present invention such analysis is performed independent from absloute concentrations of epitope variants in a single measurement from one and the same sample. This has the striking advantage that the relative concentration of epitope variants, especially of one of the variants representing old bone to the variant representing new bone, directly translates to bone metabolism or bone balance.

Preferably for each of the variants representative for old bone, βL-peptide variants, αD-peptide variants and the βD-peptide variants the relative concentration as compared to the αL-peptide variants is analysed. Most preferred is the use of a ratio between an αL-peptide variant and a βL-peptide variant.

Methods to establish standard or "normal" values are in-numerous. They depend largely on the "normals" chosen, the assay procedures applied and to a great deal on the procedures used to standardise or calibrate these assays. In the case of bone metabolism or bone balance or rate of bone loss, no internationally accepted biochemical marker is available. However, it is possible to obtain such standardisation along the following lines. The group of normals may be selected to comprise healthy adults e.g., aged 30-40 years, without any history of a bone- related complication. It is known from densitometric measurements that within such a group of individuals bone mass stays more or less constant. In biochemical terms there is bone balance and neither bone loss or bone gain is significant. In mathematical terms the rate of (bone) loss is 1 in such a group of individuals. Specific assays e.g., based on the principle of competitive immunoassays are available or can be designed to establish concentrations of these variants individually. Such individual values are used to calculate ratios of variants. Such individually obtained values are also used to establish standard or calibration curves for variant ratios. Such calibration curves usually are set up to cover at least 90% of the measuring range observed with biological samples for these values. Unknown samples can be analysed and ratios of variant extrapolated from previously or parallel established standard curves.

To establish normal or reference values usually a so-called reference group (see above) is selected. The relative concentrations of epitope variants, e.g. the αL to βL-ratio are established for this group of individuals. The average value for these samples is calculated and used to define normal bone metabolism, equivalent to bone balance (no loss and no gain) and expressed by a resorption rate ( e.g. derived from the αL to βL-ratio) equal to 1. For different clinical problems different reference groups may be used to establish reference values for methods and devices according to the present invention.

In individuals losing bone e.g., as diagnosed by densitometry, metabolic bone balance or net bone balance is negative, bone resorption rate or rate of bone loss as derived from the relative amount of bone formed devided by the relative amount of bone resorped, e.g. as assessed by the αL- to βL-ratio, is smaller than 1. In case of successful treatment with so-called anti- resorptive drugs, e.g., hormone replacement therapy, or treatment with a bisphosphonate, like Alendronate® or Bondronate®, bone balance will improve. Such improvement on bone metabolism so far finds its biochemical correlate in greatly reduced levels of bone resorption markers (Bjarnason, N. H., Christiansen C, Bone 26(6), p. 553-560, 2000). With the methods now at hand due to the present invention efficacy of treatment can be assessed by improvements on the rate of bone resorption. Effective treatment results in bone resorption rates (now better termed formation rates) of greater than 1.

It is preferred to use a method according to present invention in order to assess bone metabolism.

It is further preferred to use a method according to present invention in order to assess efficacy of treatment regimens effecting bone metabolism. As mentioned, diagnosis of osteoporosis even today suffers from severe limitations. One of these limitations is that the absolute amounts of biochemical markers even as assessed with most advanced technologies still are dependent both on bone mass as well as on formation or resorption rate. Results obtained according to the present invention do only depend on one of these two variables, the rate of bone formation or resorption. The absolute amount of bone mass present has no or only negligible influence on the ratio of variants of collagen degradation products as assessed here.

Diagnosis of osteoporosis is based on a variety of factors and even in the future will not be made based on a single measurement according to the present invention. However, use of a method according to this invention will greatly enhance the screening potential for osteoporosis or probably better said for patients having a negative bone balance who very well might be at risk to develop osteoporosis. This greatly improves on the tools available in aiding diagnosis of osteoporosis. Once identified, individuals at risk then should be carefully and closely examined by a clinician. Most importantly, if required, patients can be alerted to the disease, change their live style, or treated with anti-resorptive drugs early on and efficacy of treatment can be assessed used the procedures of this invention.

It is therefore another preferred embodiment to use a method according to present invention in diagnosis of osteoporosis.

It is another striking feature of the present invention that a large variety of sample sources may be used. Such sample may be highly variable without negative impact on results as long as enough analyte carrying the variant of an epitope recognised by the first specific binding partner is available, or, as long as measures are taken to ensure that enough sample has been made available.

Samples to be analysed may be taken from standard sources like blood, serum or plasma, but also highly variable samples like tear liquid, saliva, exudate fluid, synovial fluid, cerebrospinal fluid, or tissue extracts may be used. Usually such highly variable samples are quite difficult to handle, whereas this is not the case for methods and devices according to the present invention.

In EP 0 753 148 B1 saliva is used as source of analytes. Reference is given there to a lot of patents dealing with problems associated with saliva as a sample and with appropriate solutions to tackle some of these problems. Probably the most serious problem, when using saliva as a sample, is the fact, that saliva is not a uniform sample. Rather different concentration/dilution of this sample and therefore also of analytes contained therein will regularly be encountered. Various patents have sought to define means to correct for the relative dilution/concentration (=concentration and or dilution, respectively) of analytes in saliva samples. US 5,534,502 and PCT/WO 93/11434 describe devices for determining that a (saliva) sample collected for diagnostic purposes is in fact saliva. Both are based on demonstrating the presence of amylase activity in the sample.

EP 0753 148 B1 further improves the testing of analytes in saliva by defining an amylase cutoff-value. By defining such an amylase cut-off it is ensured that at least the minimum amount of sample required for a meaningful analysis is present. However, it will be obvious to the expert in the field that a device as described in EP 0 753 148 B1 only can produce qualitative, yes-or-no types of results. An amylase cut-off is not required when using the methods and devices according to the present invention.

Recently methods have been developed and means have been designed to collect so-called exudate fluid. The procedures and devices described in US 6,048,337 are herewith included by reference. Exudate fluid is quite variable and measures for standardisation have to be taken. Exudate is known to contain small peptides, e.g. peptides derived from collagen degradation. CDPs can easily be measured in* exudate fluid according to methods and devices of the present invention.

Most diagnostic procedures today are based on the analysis of serum, blood or plasma. Rather few parameters are diagnosed from urine despite the fact that it can be most easily obtained. The major disadvantage of urine resides in the fact that it does not represent a uniform more or less constant sample source. Instead, the concentration of analytes in urine is largely dependent on the urinary excretion rate which in turn is dependent on a lot of parameters of different origin, like fluid intake, physical activities, day-time, stress etc.. Where only qualitative results are required, e.g. for certain drugs, urine is the sample of choice.

Whenever an analyte shall be determined in a manner not dependent an urine dilution or excretion rate, special and time-consuming measures have to be taken. Either 24 hour- collections of urine have to be considered or the analyte measurement has to be corrected for sample dilution by aid of a reference molecule, e.g. like creatinine. According to the present invention the ratio of variants of epitopes on an analyte is measured. The result obtained with methods and devices according to the present invention is a relative measure of analytes and not a measure of absolute concentrations. Therefore the problems encountered due to the high variability of sample sources like saliva, exudate fluid, or urine so far are coped with.

It therefore is another preferred embodiment to use samples like saliva, exudate fluid or urine in a method according to the present invention and that no correction measures for sample volume or sample dilution are required.

High variability of analyte concentrations has been shown for all collagen degradation products. This high variation of marker molecules, e.g., within a single person or patient represents the most severe drawback and a most critical obstacle to the broad diagnostic use of CDP markers (cf. e.g., Hannon et al., (1998) Journal of Bone and Mineral Research, 13:1125-1133). High diurnal variation has been shown for any CDP-based product on the market and investigated there. The diurnal variation of urinary markers may be as high as + 50 %. Such high diurnal variation requires extremely high changes in marker levels in order to obtain clinically significant changes or in order to be of diagnostic significance.

High diurnal variation is extremely critical for all CDP-based products known. Most likely this variation is caused by high bone metabolic activity during night on the one hand and low metabolic activity of bone tissue during daytime. The absolute values of any known CDP-based marker are subject to these inevitable problems caused by normal physiology of human beings. Existing methods try to at least partially cope with the problem by standardising urine sampling, e.g. by always collecting first or second morning void urine. It has surprisingly been found that the ratio of variants, as determined according to the present invention is essentially independent of diurnal variation. This represents a big step forward and greatly enhances the applicability of any diagnostic means based on this invention.

It therefore is a preferred embodiment of the present invention to use the inventive methods and devices for measurement of urinary samples. It is also a big advantage that no care has to be taken regarding the day-time of collecting the sample.

Devices according to the dry chemistry principle are used with great advantage in combination with the methodological principles according to the present invention. Standard dry chemistry devices, e.g. test strip-like devices have recently been developed in many different arrangements to cover various different clinical indications, sample sources and analytes. Especially for urinalysis there have been many improvements and it is not possible neither necessary to cover this field in great detail here, since the skilled artisan has no problem to find the information necessary to carry out the invention in the literature pertaining to this field. The use of reagent-impregnated test strips in specific binding assays, such as immunoassays is described in the relevant patent literature. Methods, procedures and test strip materials described e.g. in GB 1589234, EP 225054, EP 183 442, EP 186 799, EP 212 603 and EP 291 194, are included herewith by reference.

In a preferred embodiment the sample is applied to one part of the device and is allowed to permeate through the other parts and material(s) of the device. The sample itself or a "solvent", in most cases water or a water-based buffer elutes all the reagents present in the device and necessary to obtain specific binding reactions. The complexes between analyte and first specific binding partner are formed in or trapped into a detection zone and detection is performed by aid of a labelled variant-specific second binding partner. This second variant-specific binding partner preferably is part of the device. It may however also be mixed to the samples or applied to the device subsequently and used as appropriate.

The first variant-specific binding partner is bound to the detection zone or capable of binding thereto and present in limited amounts. Binding is preferably direct ,e.g. by adsorption to appropriate materials, e.g., nitrocellulose or by co-valent chemical linkage. In another preferred embodiment the detection zone is coated with streptavidin and the first variant-specific binding partner is biotinylated and bound there indirectly, e.g., during analysis or when manufacturing the device.

It is preferred that the inventive device is designed to comprise a sample application area, for which many kinds of material are appropriate and may be used. This area may be also impregnated with reagents supporting the detection method.

It is preferred that the inventive device is designed to comprise a means for facilitating liquid permeation through this device. This means most easily ensures that analyte molecules pass through the detection zone. Said means may comprise the test strip material itself in dry form. Special means may be construed to work as an "absorbant sink". Such absorbant sink may comprise e.g., Whatman 3MM® chromatography paper and e.g., provides enough absorptive capacity to allow that sufficient amounts of sample passes the device and any unbound detection reagent is "washed" away from the detection zone (cf., e.g., EP 0 291 194).

In case analytes shall be quantified by state-of-the-art dry chemistry test strip devices, the volume applied to the strip has to be controlled, e.g., by applying only the required and defined volume or amount of a sample to the device, or by using devices which are designed to only take a defined volume of sample, e.g. by providing for application zones, which take a defined volume of sample, and/or by manufacturing the whole device to only analyse such a defined volume. It is only this way possible to quantify an analyte by a dry chemistry device, e. g. by aid of a test strip. However, after the analyte has been measured from a variable sample source, e.g. urine, correction measures still have to be used to correct for concentration/dilution effects of the sample investigated. This problem is obviated by the present invention. Using the procedures and devices disclosed, it is neither required to measure independently the concentrations of at least two different epitope variants on an analyte, nor is it necessary to apply exactly controlled volumes of a sample to the device.

Since the methods and devices according to the present invention work independent of assessment of absolute concentrations of epitope variants, no special precautions regarding sample volume have to be taken, it is e.g. possible to hold or dip the device into the sample, and not to control for the actual sample volume taken up by the device, as long as enough sample overall - first epitope variant in excess - is applied.

A preferred embodiment of the present invention is a device according to the dry chemistry principle for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants, comprising at least a sample application area, a means facilitating uni-directional liquid flow and a detection zone, characterised in that the first variant- specific binding partner is present in limited amount and bound or capable of binding to the detection zone of said device.

In order to reduce handling steps it is preferred to include the second variant-specific binding partner into the device. Various ways are possible. Most preferred the second variant-specific binding partner, by aid of which variant-specific detection is made, is present in high amounts, e.g. in the sample application area. Most preferred incorporation is made in such ways to facilitate continuos or slow release of this reagent. Such ways of incorporation are e.g. described in EP 0 186 799.

A preferred embodiment according to the present invention therefore is the above described inventive device which is further characterised in that the second variant-specific binding partner present in excess is incorporated into the device.

In many cases it is desirable to not simply believe that enough analyte has been applied but rather to make sure that enough analyte is available. With dry chemistry devices this can easily be accomplished by including a safety zone into such a device. The safety area or zone is manufactured to indicate to the user that enough sample has been applied and that the result of the reaction may be evaluated, measured or read. The safety zone is construed into the device at a location reached by the sample liquid or reaction mixture after having passed the detection zone. In a preferred embodiment the safety zone makes use of a binding partner exhibiting essential identical specificity as compared to the first variant-specific binding partner, most preferred of the first binding partner itself. Many designs are possible utilising the specific binding between the first binding partner and the first epitope variant. In a preferred embodiment the variant-specific first binding partner is labelled by visually clearly seen means or labels and bound via the first epitope variant or by an analogue thereof to the device. Analogues facilitating the release of such bound (labelled) first variant-specific reagent are preferred. Once the overflow of analyte carrying the first epitope variant reaches the safety zone, the labelled binding partner is released, thus indicating that this part of the reaction (the formation of complexes between first binding partner and analyte) is completed. This step can be evaluated by eye or appropriate measurement devices.

It therefore is a preferred embodiment of the present invention that the inventive device is designed to also contain a safety zone.

It is good practice to include a positive reagent control area into chromatographic or test strip devices, such doing is also preferred for the devices described here. It is preferred to include a means to prove that the detection reagent works as required. In case of the present invention the epitope variant to which the labelled second specific binding partner (the detection reagent) binds is incorporated into the device by appropriate means at a location after liquid has passed the detection zone. Labelled detection reagent will bind there indicating that reaction is complete and that this reagent was properly working. The ratio of epitope variants in the sample is correlated to the ratio of epitope variants bound in the detection zone. This ratio is correlated to the fraction of analytes comprising the second epitope-variants trapped in the detection zone and is indicated by the labelled second variant- specific binding partner bound there. Result of measurement is read by eye or appropriate measurement devices are used to quantify any label bound there.

Of course the inventive device has to be targeted to specific analytes of interest. As described an discussed above degradation products comprising the cross-linked collagen telo-peptides are highly important analytes. In a preferred embodiment the device according to the present invention is further characterised in that the variant-specific binding partners recognise collagen degradation products, especially the ones termed and known as NTX or CTX.

Assessment of bone metabolism even today is rather difficult. However screening in the relevant part of the population for those people at risk to loose bone or to develop osteoporosis would be very important and lead to big long term benefits both for patients as well as for our health care systems.

Many of the technical complications with state of the art procedures to assess bone metabolism or bone balance have already been described. With few exceptions these methods and procedures require that a "potential" patient visits a physician, that blood or urine is taken, that the samples are analysed in a central lab and that the patient re-visits the physician. Alternatively densitometric or radiographic procedures may be used. A high number of non- patients will be amongst the people investigated and very high costs will be incurred.

Pre-selection of "candidate" patients is the method of choice. From the procedures known in the art and commercially available the device described by Blatt et al., supra, may be considered most advanced. J. Blatt et al., describe such a device which is "A Miniaturized, Self-contained, Single-use, Disposable Assay Device for the Quantitative Determination of the Bone Resorption Marker NTX in Urine". This (single-use!) handhold device contains all the reagents as well as equipment (including battery) necessary to measure both the signals developed by the NTX- specific reaction as well as by the creatinine determination, to calculate the ratio of NTX/creatinine and to display the rationed result. This is sophisticated but technically most advanced. However, like with other creatinine-corrected CDP markers it can not cope for influences due to body mass or even more critical due to diurnal variation. Quite different the devices according to the present invention are technically much less demanding - which in most cases is known to translate to less complications. They provide results of only a single measurement, are easy to handle and comparatively cheap to manufacture and nonetheless provide an attractive means to assess bone balance rather independent of actual bone mass and most strikingly are not influenced by diurnal variation.

In a preferred embodiment the devices according to the present invention are used to assess bone metabolism, and especially bone balance.

Various treatment regimens targeting at bone metabolism, mainly anti-resorptive therapies like hormone replacement therapy, vitamin D, and bisphosphonates are available. Some of them have been proven to be quite effective in treatment of post-menopausal osteoporosis. It is, however, one of the big problems in the field that efficacay of treatment - as assessed by bone mineral density - is first clearly seen after one to three years of therapy. Very recently it could be demonstrated that some biochemical markers might be useful to assess efficacy of treatment in an individual patient (Bjarnason, N. H., Christiansen C, Bone 26(6), p. 553-560, 2000). Such measurement being possible a few weeks to months after start of treatment. Such diagnostic means are of great advantage. They allow to diagnose non-compliance and patients failing to respond to therapy early on. The same holds true for the methods and devices of the present invention. The inventive device is cheap easy to use and does not require a central laboratory. In a further preferred embodiment a device according to the present invention is therfore used to assess efficacy of treatment regimens effecting bone metabolism.

Diagnosis of osteoporosis is costly and difficult. However, and even more important osteoporosis is extremely under-diagnosed. It is estimated that only about 20% of people suffering from osteoporosis actually are diagnosed as osteoporotic patients and even less are treated with one of the available efficient drugs. Not the least problem in the field of osteoporosis is the availability of comparatively cheap and reliable means to assess lots of people, e.g. in screening programmes. By using the inventive devices it is possible to overcome these problems, to assess rate of bone loss in lots of "potential" patients, e.g. in a whole population cohort, and to select for those patients having an increased rate of bone loss. Such means will greatly aid in the overall diagnosis of osteoporosis. ln a preferred embodiment the methods and devices according to the present invention are used in diagnosis of osteoporosis.

Examples:

Example 1 : Production of antibodies with specificity for collagen degradation products

Antibodies to variants of the collagen telo-peptide sequences are obtained according to procedures described in patent application PCT/EP 97/04372. The basic sequences used are:

N-telo-peptide sequence of type I collagen (α1): Asp-Glu-Lys-Ser-Thr-Gly-Gly N-telo-peptide sequence of type I collagen (α2): Gln-Tyr-Asp^*-Gly-Lys-Gly-Val-Gly C-telo-peptide sequence of type I collagen (α1): Glu-Lys-AlarHis-Asp*-Gly-Gly-Arg * this peptide bond is subject to isomerisation and the βL-, βD- and αD-forms are produced and used as described in PCT/EP 97/04372.

Synthetic petides are coupled to a carrier molecule, to immunise laboratory animals. Screening for appropriate (variant-)specific antibodies antibodies is performed in a competitive immunoassay format using free peptides of the investigated variants as competing agents. Furthermore antibodies are selected to be capable of forming a sandwich with the target analyte, comprising the desired variants.

Example 2: Modification of variant-specific antibodies

Antibodies derived from clones 1103 and F12 have been used. Most of the reagents referenced below are part of commercially available kits (one-step ELISA for serum-CrossLaps® by Osteometer, Copenhagen, Denmark; Serum CrossLaps Elecsys® by Roche Diagnostics Mannheim, Germany). Appropriate dilutions have to be tested.

a) Fragmention of antibodies

Where mentioned, purified IgG of either clone has been used to generate F(ab) or F(ab^')₂- fragments. Such fragmentation has been performed using standard enzymatic digestion methods employing pepsin or pa^'pain, respectively. b) Biotinylation

Biotin-labelling is performed using biotinyl-N-hydroxysuccinimide ester as coupling reagent. Coupling is carried out at protein concentrations above 10mg/ml in 100 mM carbonate buffer at pH 8.5. BNHS is dissolved in N,N-dimethylformamide (DMF) immediately prior to use and mixed with the antibody solution in a molar ratio of approximately 3:1. Reaction is performed for 4 hours or overnight. Purification is performed by dialysis.

c) Peroxidase-labelling

Labelling of IgG or the above described fragments is performed according to the sodium periodate (Nal0₄) method, as described on page 236 in Tijssen, 1995, supra. Conjugation mixture is purified and fractionated using a Sephadex® G-25-column. Optimal fractions, i.e. fractions comprising conjugate yielding high specific signal and very low back-ground recativity are selected and optimal dilution (best signal to noise ratio) for conducting the assay is determined.

d) Labelling of antibody or fragments thereof with colloidal gold particles

Gold sol, was prepared by the hydroxylamine mediated reduction of tetrachloroauric acid in water onto seed gold particles. This procedure is described in the literature: Turkevich, J. et al., Discussions of the Faraday Society, No. 11, p 55-74. Gold sol is coated with aminodextran of 40 000 D and antibody is coupled to the dextran-coated gold particles with (1-ethyl-3(3- dimethylaminopropyl)-carbodiimid (EDAC).

Example 3: Determination of analyte variants in an ELISA assay ELISA-plates pre-coated with streptavidin (Micro-Coat, Bemried, Germany) are used as solid phase. Plates are incubated with 100 μl of a solution containing 50 ng/ml of biotinylated F(ab)- fragments of clone 1103 in PBS/Tween-buffer (phospahate buffered saline, pH 7.5 with 0.05 % Tween® 20) per well. After 1 hour incubation at room temperature wells are washed three times with 300 μl PBS/Tween per well. Second morning void urine samples (quite concentrated as can be told from their appearance) are diluted 1 to 10 in PBS/Tween and 100 μl per well incubated for one hour. After three washes with PBS/Tween.100 μl of detection reagent (F12- peroxidase) is incubated for 30 min or one hour. Detection is performed using 3,3',5,5^'- tetramethylbenzidine (TMB) as substrate and read by standard ELISA-reader equipment. Optical reading for ten urine samples taken from healthy adults aged 30 to 40 all yield optical densities in a very narrow range, averaging to 400 mE.

A) Results with clinical samples:

B) Results within one patient during the course of one day and expressed as percentage of the value obtained from mixing 2ml each of urine from any time-point.

These data clearly prove that the relative concentration of the measured CTX-variant (comprising twice the β-variant of the 8AA-epitope) is over the full day rather constant. The ratio of CTX (CTX = urinary β-CrossLaps as measured with the Osteometer kit) CTX to creatinine exhibits the well-known diurnal variation, even after correction for volume or dilution effects by measurement of the respective creatinine concentration in each urine sample.

Example 4: Test strip device

The key features of this invention have been tested in prototype devices using nitrocellulose sheets. Nitrocellulose sheets are cut into small strips (about 4mm by 50mm). F(ab)-fragments of antibody 1103 are directly coated onto the nitrocellulose strip in form of a band or line. This comprises the detection zone. All remaining protein binding sites are blocked by P BS containing 1 % BSA. Urine is applied left to the detection zone and liquid flow through the device facilitated by placing the strip right to the detection zone on filter paper, readily accepting the flow through liquid. Once at least about 20 μl of urine have passed the device, this part of the reaction is stopped and the device washed. Detection reagent (gold-labelled IgG of clone F12) is then either dropped onto the detection zone or the procedure already used for urine is performed again. Non-bound detection reagent is removed by washing and evaluation performed by eye.

Urine samples from a patient with active Paget^'s disease, a healthy control, and a patient with active primary hyperparathyroidism (PHPT)are tested. When comparing results of the patient samples to the result obtained with urine from the healthy control, it is obvious, that the signal is much stronger with the PHPT-urine and much decreased using the urine collected from the patient with active Paget's disease, respetively.

With the methods and devices described and the examples given at hand the skilled artisan will be able to vary the concepts according to his special requirements without departing from the spirit of this invention. E.g., the set-up of dry chemistry devices may be varied to a great extend. Thus it should be understood that although the present invention has been disclosed in preferred embodiments, modifications and variations of the concepts disclosed herein may be resorted by those skilled in the art, and that such modifications and variations are considered to be within the scope of this invention as defined by the appended claims.

Claims

I claim:

1. Method for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants, characterised in that,

(a) a first variant-specific binding partner

(b)a sample containing said first variant of said epitope in excess as compared to said first binding partner are incubated under conditions allowing for binding between said epitope variant and said first binding partner

(d)a detectably labelled second variant-specific binding partner is used to detect the fraction of complexes (c) carrying the epitope variant recognised by said second binding partner

(e)the signal generated (d) is correlated to a calibration or standard curve.

2. Method for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants, characterised in that, (a) a first variant-specific binding partner

(b)a sample suspected of containing said first variant of said epitope are incubated under conditions allowing for binding between said epitope variant and said first binding partner

(c) measures are taken to ensure that said first variant of said epitope is available in excess as compared to said first variant-specific binding partner

(d) complexes between said first binding partner and said analyte are formed wherein the ratio of variants in said complexes correlates to the ratio of said variants in said sample (e)a detectably labelled second variant-specific binding partner is used to detect the fraction of complexes (d) carrying the epitope variant recognised by said second binding partner

(f) the signal generated (e) is correlated to a calibration or standard curve.

3. The method according to claim 1 or 2 characterised in that the first and the second variant-specific binding partner recognise different variants of said epitope

4. The method according to claim 1 or 2 characterised in that the first and the second variant-specific binding partner recognise the same variant of said epitope

5. The method according to one of claims 1 to 4 characterised in that the first and the second variant-specific binding partner are antibodies

6. The method according to one of claims 1 to 4 characterised in that the first and the second variant-specific binding partner are monoclonal antibodies

7. The method according to one of claims 1 to 6 characterised in that said analyte is a collagen degradation product

8. The method according to one of claims 1 to 7 characterised in that the collagen degradation product is an N-terminal telo-peptide

9. The method according to one of claims 1 to 7 characterised in that the collagen degradation product is a C-terminal telo-peptide

10. The method according to one of claims 1 to 9 characterised in that it is used to assess bone metabolism

11. The method according to one of claims 1 to 10 characterised in that it is used to assess efficacy of treatment regimens effecting bone metabolism

12. The method according to one of claims 1 to 9 characterised in that it is used in diagnosis of osteoporosis

13. The method according to one of claims 1 to 12 characterised in that saliva, exudate fluid or urine are used as samples

14. Device for the assessment of the ratio of variants of an epitope on an analyte molecule, comprising at least two variants of said epitope, by aid of a first and a second specific binding partner for said epitope variants, comprising at least a sample application area, a means facilitating uni-directional liquid flow and a detection zone, characterised in that the first variant-specific binding partner is present in limited amount and bound or capable of binding to the detection zone of said device.

15. The device according to claim 14 characterised in that the second variant-specific binding partner present in excess is incorporated into the device

16. The device according one of claims 14 or 15 characterised in that the device also contains a safety zone

17. The device according to claims 14 to 16 characterised in that the variant-specific binding partners recognise collagen degradation products

18. Use of a device according to one of claims 14 to 17 to assess bone balance

19. Use of a device according to one of claims 14 to 18 to assess efficacy of treatment regimens effecting bone metabolism

20. Use of a device according to one of claims 14 to 18 in diagnosis of osteoporosis